Adversarial idea evaluation for Claude Code - kill bad ideas before they eat money, time, and attention. Every verdict compounds into a persistent, auditable decision ledger.

The hand-made edge: direction-first design taste for Claude, on the canvas and in code.

A reusable Agent Skill for turning unstructured knowledge into structured, searchable personal knowledge.

12 free Claude Skills for small business — invoice chasing, review replies, local SEO audits, landing page teardowns, and more. Real SOPs, not prompts.

Claude/agent skill: fair coaching rates by role, territory, and coach seniority — ICF 2025 + CEE market anchors, with red flags before you pay

A Claude-powered trip planning workspace template — skills for route planning, food/stay scouting, itinerary building, and mid-trip replanning

A Claude Code system that keeps competitive battle cards current on their own: cited claims, watched sources, auto-rebuilt when a competitor moves.

Claude Code skill: find, verify, and prioritize B2B leads in any city — evidence quotes, adversarial verification, 21 vertical-by-city presets, Excel-ready output

A collection of reusable skills for Claude and other LLMs. Skills can be imported into Claude and other LLMs that use projects, or copy/pasted into any chatbot window.

🧭 深度讨论协作 Agent Skill - 维护线索地图,陪你想清模糊想法。三面身份(讨论者/导游/播种者)+ 七种关系 + 收线沉淀。

Continuous fullstack delivery loops — orchestrates frontend, backend, and quality subagents (behaviour drivers, engineers, UI/UX specialist, code/security reviewers, architects) in a test → diagnose → fix → review → secure → re-test cycle until the work is production-ready

Claude/agent skill: scrape DuckDuckGo search results into structured JSON. Installs via npx skills add.

Claude/agent skill: turn App Store reviews into ASO and product feedback insight. Installs via npx skills add.

Claude/agent skill: current gas and fuel prices as structured data. Installs via npx skills add.

OWASP Top 10 (2025) security audit plugin for Claude Code and Cursor. Stack-aware scanning that produces a complete vulnerability report.

Training wheels for AI assisted coding. Seam aware protocols that make coding agents investigate first, gate irreversible actions, and prove their claims.

I got tired of Claude jumping to sloppy, hastily-formed conclusions when I was processing batches of semi-structured research data: interviews, transcripts, surveys, observation notes, etc. This skill teaches Claude to analyze the inputs ("what's in the data?") BEFORE doing synthesis or ideation ("what does it mean?").

Novelty-claim guardrail for Claude Code & Grok — provenance + measured-delta discipline before you claim a first.

Agent skill (SKILL.md) that makes Claude and other coding agents write prose without AI tells — puffery, fake analysis, 'not just X but Y', overused vocabulary, formatting spam.

Instead of writing comments, your coding agent asks you to explain the code it just wrote, then grades the answer and awards XP. Retrieval practice for AI-assisted development.

Local session memory for Claude Code and Codex—capture context, decisions, artifact references, and next state in Markdown and JSON.

Agent skills for authoring HelpNest help-center content — installable via skills.sh (npx skills add).

Claude/agent skill: scheduled hotel rate snapshots for rate drop alerts and competitor watching. Installs via npx skills add.

Claude/agent skill: bulk export job listings to CSV, Excel, Sheets, or JSON with quotable per-row cost. Installs via npx skills add.

Claude/agent skill: bulk Yandex SERP collection with predictable per-result billing. Installs via npx skills add.

Claude/agent skill: turn Google Jobs into a job search API feed for apps and AI agents. Installs via npx skills add.

Claude skill for auditing and refreshing existing blog posts (SEO + GEO).

AI Ready Software Template for Enterprise Multi-Model AI Context Architecture

一条命令去掉口播录音的句间气口/换气声，节奏立刻紧凑 · Remove breath gaps between sentences in voiceovers. ffmpeg-based, also a Claude Code skill.

上线前测试审计 Skill：用于发布前检查、回归测试规划、安全审计和 Go/No-Go 上线决策。

A skill that lets AI agents draft friction reports about other skills, for human review and upstream filing

Komentarze do raportów marketingowych jak od specjalisty, nie generatora. | Client-facing marketing report commentary in Polish, without the AI filler.

Ready-to-install agent skills, each wrapping a nanoodle visual AI workflow — poster, photo restyle, jingle, narrated poem, video teaser

Claude Code skill: deterministic subagent budgeting — token forecasts per worker class, model-tier recommendation, worker contracts with escalation triggers

Turn any topic into an expert-level AI skill — for Claude, ChatGPT, Gemini, or any LLM. A meta-skill that researches a domain and writes the skill for you.

Public base configs for Claude Code (skills, commands, agents) -- a working ccs payload repo. Content arriving via curation.

Claude skill that strips machine tells from writing: banned AI vocabulary, structure symmetry, burstiness, hedging density. Authentic writing, no detector tricks.

Claude Code Agent Skills for building, red-teaming and tuning agentic RL evaluation environments — a four-skill pattern (guardian, validation-debugger, score-tuner, iteration-loop) plus a 24-point adversarial reviewer.

Root-level Claude Code skill for Tree Ring Memory: local-first recall, capture, audit, consolidation, and forgetting workflows for AI agents.

Claude skill: 未來學研究方法分析器 — turn qualitative & trend data into rigorous futures-studies analysis (Six Pillars, Delphi, CLA, scenario planning, futures triangle/wheel, backcasting and more)

Claude Code skill that makes any model (Sonnet 5 / Opus 4.8) operate with Fable-class discipline - task-adaptive practice routing, verify loops, agent orchestration, adversarial self-checks, outcome-first communication. Install into .claude/skills/ and invoke with /fable-mode.

Claude Code skill for safe, fast Cloudflare D1 + R2 operations on Windows - REST-based, immune to PowerShell BOM/encoding traps

Claude skill: AI 角色扮演練習夥伴 — rehearse difficult conversations with AI playing any persona, with structured feedback

Portable Agent Skills (SKILL.md) for Claude Code and compatible CLIs

Agent skills for Claude Code — skill-selector (choosing skills) + skillviz (visualizing them)

Reusable Claude Skill and LLM agent framework inspired by Fable 5 reasoning workflows, prompt orchestration, tool simulation, validation, and project-level automation.

Offload Claude Code's heavy coding to Codex; keep judgment and verification on Claude.

Production Claude Skills from a working fractional CMO practice: client sweeps, critique layers, decision councils, Meta engines

iOS Liquid Glass Skill for Swift and React Native

把飞书妙记的全部录音转写拉到本地 Markdown(含 AI 总结)——不建应用、不申请 API 权限、不用管理员审批，浏览器登录一次即可。Pull all your Feishu Minutes transcripts into local Markdown with AI summaries.

Decision records, templates, and practice scenarios for the call behind a release, launch, or production milestone — captured before the outcome is known, closed out after it is.

Two interoperable Claude Agent Skills: evidence-grounded product-market strategy and conversion copy, with claim traceability and no fabricated proof.

Security audit skill for Claude Code — scans other skills for prompt injection, data exfiltration, and privilege escalation before you install them.

A Claude Skill that turns Claude into a senior freelance & creator business strategist — niche selection, pricing, client acquisition, AI automation, WordPress trends. No yes-man advice, just ranked, practical strategy.

Agent skill that systematically guides AI to analyze requirements and design comprehensive, traceable end-to-end test scenarios before scripting

Claude/agent skill: monitor news coverage and brand mentions on a schedule. Installs via npx skills add.

Claude/agent skill: scheduled fuel price snapshots for tracking changes. Installs via npx skills add.

Claude Code skill — CZ CELLxGENE Census API & Discover REST reference for querying single-cell RNA data. 1.18.0-verified.

A nine-stage art-directed Claude Code workflow: commit direction, generate original media, review mechanics independently, deploy, and leave receipts.